A Terrible Argument Against Congestion Pricing
Via Ryan Avent, David Owen says that congestion pricing would be bad for the environment because traffic jams inspire people to take transit instead of driving.
This is, in my view, a very silly line of argument. It’s probably true that you could construct a model of a situation in which congestion pricing increases the net quantity of driving. But if that situation exists, and you want to change it, then there are lots of good policy options available. You could use the revenue from congestion pricing to finance more attractive transit options. Or you could take advantage of the reduce congestion to start taking lanes away from private automobiles and building bike and bus lanes. You could do all kinds of things.
The main point I would make is that the issue of whether or not you should congestion-price roadways is more-or-less at right angles with the question of how much your public infrastructure should promote driving versus cycling or transit or walking or anything else. The point of congestion-pricing is that the most efficient way to manage the scarce resource of space on crowded streets during peak hours is via a congestion price. That’s true no matter how much or how little driving you’re hoping to see. If you want people to drive less, the thing to do is to build narrower roads and invest in transit and bike infrastructure. If you want people to drive more, the thing to do is to build narrower roads and be stingy on transportation alternatives. But either way if you want to avoid productivity-killing traffic jams you ought to charge people for driving at peak hours.
38 Responses to “A Terrible Argument Against Congestion Pricing”
Register
RSSLogin
Topic Cloud
ArchivesBlog RollAbout Matt YglesiasContact Me
Donate
October 13th, 2009 at 5:52 pm
Good thing make not bad enough to push good.
Tarzan algebra.
October 13th, 2009 at 6:00 pm
I agree, but would have used the word “orthogonal”… because fancy geometry terms are hot.
October 13th, 2009 at 6:01 pm
wait what? How is David Owen’s argument be even remotely plausible? He’s saying that by charging people to drive possibly we end up with more cars on the road? How? If you charge people to drive in say, NYC, they will drive in the city less and they will car pool.
If they’re driving less, then either they are sitting at home not going anywhere (good for the environment), going somewhere else (neutral for the environment) or using public transportation (good for the environment).
If they are car pooling, that still won’t increase the total number cars on the road. You’ve still raised the cost of driving for everyone in the car. Plus if roads ended up packed with car poolers, then you’d just raise the price of driving further.
Is there really any way for what David wrote to be correct?
October 13th, 2009 at 6:07 pm
David Owen’s new book raises an excellent point that density is a good thing, but some of his side points are just absurd — including this one about congestion charging.
I think Owen thinks the point of the congestion charge is to make traffic better — but the real point of it is, as Matt points out, is to reduce traffic AND to provide more non-motorized options and transit (including to areas not served by it).
October 13th, 2009 at 6:11 pm
A modestly related irony is that if congestion pricing were working well, there would seem to be little need for it.
October 13th, 2009 at 6:13 pm
It’s probably true that you could construct a model of a situation in which congestion pricing increases the net quantity of driving.
I’m not sure how you’d do that, unless you had some crazy congestion response to adding an additional car (or people with extreme time-money tradeoffs I suppose). Short of assuming that traffic speeds actually increase when you add cars to the road, I can’t see how to make it work, even hypothetically.
October 13th, 2009 at 6:18 pm
No, the point of congestion charging is to commoditize a public good, reduce mobility, apply a regressive taxation system and generally make peoples’ lives just a little bit more miserable as a mechanism for smoothing out the commutes of rich people.
October 13th, 2009 at 6:29 pm
Anthony – the idea that congestion pricing meant to smooth out the commute of rich people at the expense of others is absurd.
Revenue from congestion pricing leads to better transit access in transit poor neighborhoods. Those car commuters will switch from driving and parking in the city to transit, which would likely be much less expensive and just as fast. And for the many poorer people that already use transit and ride bikes, facilities will be make better for them.
Unless of course you are a government worker who has their parking subsidized by the rest of us taxpayers.
October 13th, 2009 at 6:29 pm
Ding ding ding. When people talk about liberal elitists, this is what they are talking about. A hugely stupid idea.
October 13th, 2009 at 6:42 pm
Serious consideration of congestion pricing is so far out of step with popular political sentiment that wonks who suggest it might very well be more interested in policy than getting anything done.
Traffic in much of the country — and I write from Seattle which has TERRIBLE traffic for a non-New York situation — is simply not bad enough to tilt popular will to serious consideration of pricing roads. What goes along with congestion pricing MUST be vast investment in mass transit. Such investment is nowhere on the horizon. By and large people like it the way it is and do NOT want the sort of spatial structure in which congestion pricing might work.
In other words, Manhattan is sui generis. OK maybe throw in downtown San Francisco. Every place else if you charge for going into a downtown then you will simply shift trips to a non-charging location. Ultimately you have to charge for street use throughout a whole region and that is not going to happen. Period. Full stop.
October 13th, 2009 at 6:46 pm
“Nobody goes there anymore. It’s too crowded.” — Yogi Berra
October 13th, 2009 at 6:47 pm
Um, the whole mechanism through which congestion pricing reduces traffic congestion is by changing the incentives so that PEOPLE CHOOSE TO DRIVE LESS. If the congestion pricing is working at all, there are fewer cars on the road and hence less gas burned.
October 13th, 2009 at 7:25 pm
Um, the whole mechanism through which congestion pricing reduces traffic congestion is by changing the incentives so that PEOPLE CHOOSE TO DRIVE LESS. If the congestion pricing is working at all, there are fewer cars on the road and hence less gas burned.
Um, nonsense. Congestion is not the same thing as driving. Congestion pricing is an incentive to drive less on roads and at times where there is congestion, not an incentive to drive less overall. Possible responses to congestion pricing include taking a different route, driving at a different time and substituting a different destination. The long term effect of congestion pricing may be to accelerate the decentralization of jobs and people. If you live in the New York suburbs and drive to work in Manhattan, a congestion charge on cars entering Manhattan would give you an incentive to look for a job outside Manhattan, and would give employers an incentive to move jobs outside Manhattan.
October 13th, 2009 at 7:26 pm
#10: Serious consideration of congestion pricing is so far out of step with popular political sentiment that wonks who suggest it might very well be more interested in policy than getting anything done.
http://en.wikipedia.org/wiki/Overton_window
October 13th, 2009 at 7:34 pm
No, the point of congestion charging is to commoditize a public good. . .
This is simply false. Public goods are non-rival (one person’s use of the good does not diminish others’ ability to use it) and non-excludable. Being able to drive without congestion on public roads is rival and non-excludable, and thus a common good rather than a public good. Common goods are generally overused. The way to prevent overuse, which leads to worse results (congestion, and thus slower driving for everyone, not to mention the externalities of the added pollution), is to find a way to make the good excludable. One way of doing this is to put a price on it. Hence congestion pricing.
October 13th, 2009 at 7:37 pm
At $4.50 and $5.00/gallon in gas, then congestion pricing would increase traffic. If the driver can guarantee arrival times with minimal delay, then the $8 in congestion fees cover a good deal of idling gas at $4.50/gallon.
Remember, the typical hour spent stuck in traffic costs about $20 to the consumer in lost time.
October 13th, 2009 at 7:55 pm
At $4.50 and $5.00/gallon in gas, then congestion pricing would increase traffic. If the driver can guarantee arrival times with minimal delay, then the $8 in congestion fees cover a good deal of idling gas at $4.50/gallon.
You seem to be making a “nobody goes there anymore, it’s too crowded” argument. If congestion pricing leads to minimal delay, than that means congestion pricing reduces the traffic, unless you have some weird phenomenon where adding cars to an already oversaturated road increases the speed at which people drive. If you want to have smooth driving, then you set the fee at whatever level is necessary to reduce the number of drivers to the point where the roads are no longer congested.
October 13th, 2009 at 8:00 pm
I wonder whether it makes a lot more sense to increase the price of parking in congested areas. This can be done more finely than creating an abrupt border. It also doesn’t require any infrastructure changes and it doesn’t present any privacy concerns either.
October 13th, 2009 at 8:10 pm
The “congestion charge is regressive” argument really only works in a place where driving is an absolute necessity. If we’re talking about Manhattan where most driving really is an unnecessary luxury there’s nothing wrong with charging for it. Rich people more or less by definition get more luxuries than poor people, and that’s not really some kind of moral outrage.
October 13th, 2009 at 8:16 pm
“The long term effect of congestion pricing may be to accelerate the decentralization of jobs and people.”
This is not only correct in theory, it is correct in practice. A fine example can be found in Chicago. In Chicago, the “reverse” commute of commuting out of the city towards corporate campuses in the suburbs is just as bad as the “traditional” commute in the other direction. There are a lot of reasons for this , not least of which is municipal tax policy, but the fact that parking in downtown Chicago costs an absolute fortune is definitely a factor. A lot of seemingly pro-urbanist ideas are designed to get people to live in the city, but have the perverse effect of moving jobs out of the city, thus creating the same traffic congestion problem we have now, just in reverse.
October 13th, 2009 at 8:21 pm
Addendum to #18:
I guess you’d have to have a way of enforcing entrance- and exit-time-dependent fees for parking structures. Maybe the state could have cameras monitoring the number but not the identity of cars going into the structure to discourage parking-structure owners from cheating on their taxes. This wouldn’t infringe on the privacy of drivers.
Street parking in congested areas is negligibly ever unmetered, right? Or they could just meter it, and allow people who live there with permits to park for free or for a monthly or yearly fee, if they want to allow street parking for residents.
October 13th, 2009 at 8:53 pm
Remember that light synchronization comes with congestion pricing. Thus, at a precise time, all the cars parking in one large garage can simultaneously enter the zone, have all the lights aligned, and get cars the garage with little or no waiting. This is the solution discussed regarding Manhattan.
With schedules, and especially with communications, the commute traffic in Manhattan can almost eliminate any idling delays at all. Hence the system can move far more cars with less gasoline burned up.
Real time communications also allow us to channelize a path for BRT, for example. Signaling cars to yield the right of way. Thus, with no additional cost, we get fast BRT, minimal idling time, and off peak scheduling of low priority freight.
Thus, with a modest addition of digital communications, we get a large gain in oil efficiency, probably close to 30% or 40%. Doing this in all the major metropolisis would “save” about $100 billion/year in oil imports. Yet 20 million cars with an additional $50 in digital processing costs only a billion. A huge payoff.
You get an additional pay off if you listen to the insurance companies and Ford Motor. The additional electronics can be upgraded with $400/car, adding collision and lane warnings. Each car would get about 1/3 off the insurance cost, a savings that will pay off the electronics in about six months.
And, finally another advantage. Intelligent traffic zones, with communications, makes the micro electrics much safer. Potential buyers will discover the micro car is safe in intelligent traffic zones, and we get another huge savings in the energy bill.
The government has to allow this to happen. Currently Ray LaHood (DOR secretary) has no clue, and he thinks that we are going to get up and running again with the cars we have. But oil is going to hit $100 in the ext step, and Obama is making us wait for the transportation redesign.
October 13th, 2009 at 8:58 pm
I would also like for Matt to more explicitly acknowledge that substituting congestion pricing for queue-rationing is in fact going to be regressive as far as that part of the scheme is concerned. That doesn’t necessarily mean it is a bad idea, in part because you can turn around and use the revenues to offset the regressive effects, and in part because the externalities of congestion mean this isn’t a zero-sum game. But noting the initial regressive effect is important when figuring out how to design the overall policy.
By the way, swapping congestion pricing for queue-rationing is likely to favor, not disfavor, use by people who are deriving a pecuniary benefit from that use, such as commuters.
October 13th, 2009 at 10:23 pm
I’m not sure how you’d do that, unless you had some crazy congestion response to adding an additional car (or people with extreme time-money tradeoffs I suppose). Short of assuming that traffic speeds actually increase when you add cars to the road, I can’t see how to make it work, even hypothetically.
Really? Seems pretty clear to me. An obvious example would be a congested urban freeway. If you impose a congestion charge, some number of freeway users might switch to using slower surface streets to avoid the charge. They’d be trading time for money. Their new trip by surface street might also consume more gas than their old trip by freeway.
Revenue from congestion pricing leads to better transit access in transit poor neighborhoods.
How do you know? The government might use the revenues from the congestion charge to expand road capacity. In fact, in most places that is probably a more likely use of the money than using it to expand transit. The road user lobby is far larger and more politically involved than the transit user lobby.
Also, congestion pricing wouldn’t necessarily involve an increase in government revenues, anyway. It could be a revenue-neutral credit-trading mechanism, similar to a cap-and-trade mechanism for limiting pollution. The government would determine the uncongested capacity of the road, then issue road-use credits equal to that capacity to registered vehicle owners living or working within a certain distance of the road. Credits could then be bought or sold on an electronic market.
October 13th, 2009 at 10:42 pm
The hypothetical reason congestion pricing would add more cars is that the car traffic become synchronous. Driving patterns would dis-aggregate, high priority commuters at peak hours, low priority would go earlier. Shoppers and freight haulers would price in the least congested pricing. Time shifting of traffic would tend to put more people on the road over a full traffic day spreading out the peaks.
If you followed the intelligent traffic plan to its completion, then we would likely end up with twice the amount of freight and people moving about because transportation efficiency would cause a dramatic jump in economic activity.
October 13th, 2009 at 11:10 pm
I can’t believe I’m the first one to notice the enormous chair in the middle of that photograph. Whiskey Tango Foxtrot?
October 14th, 2009 at 6:59 am
“The most efficient way to manage the scarce resource of space on crowded streets during peak hours is via a congestion price.”
Maybe. But only if your larger goal is to reducxe driving. It’s fine if that’s the case. But there are other responses with regards to scarce public goods.
For instance, let’s say we build a park in a city. it gets so crowded that nobody can play frisbee or walk their dog there. One response, I guess, would be to put up a gate and charge people five dollars to enter the park. This would clearly reduce the crowding there. But it would also have all kinds of bad impacts, like poor people couldn’t use the park.
Another option is TO BUILD ANOTHER PARK. In fact, I suspect this is the option liberals would support in this case. Or… let’s say there is a government-rubn soup kitchen that keeps running out of soup. Is the most efficient response to charge money for the soup? Is the best solution for libraries in a financial pinch to charge people for books or cards?
Again, thius is not to say that congestion pricing is a necessarily a bad idea. But clearly, liberals and progressives would have serious objections to using market mechanisms to ration other public goods. the reason they support it in this case is that they don’t like the public good in question. It has nothing to do with “efficiency.”
On the flip side, conservatives out to support it for similar reasons, but don’t.
So we appear to have a case in which liberals and conservatives are being a bit elusive about their motivations. Shocking!
October 14th, 2009 at 7:18 am
The hypothetical reason congestion pricing would add more cars is that the car traffic become synchronous. Driving patterns would dis-aggregate, high priority commuters at peak hours, low priority would go earlier.
Queue rationing should have the same effect. The fact that it doesn’t shows that there is some fundamental underlying reason why transportation patterns aggregate in the way that they do.
October 14th, 2009 at 7:27 am
But clearly, liberals and progressives would have serious objections to using market mechanisms to ration other public goods.
Yes, clearly. That is why the leading issue today among liberals and progressives is that many of our national parks charge fees. Or that you need a stamp to mail a letter with the USPS. Or that the U.S. Courts charge various fees. And so on.
The fact is that “too much use?/just provide more!” only works as a general strategy in a world of unlimited resources. In a world of scarce resources, you often have to ration in one way or another, and we can and do choose price rationing for publicly-provided goods in many other cases besides roads.
October 14th, 2009 at 7:43 am
I can’t believe I’m the first one to notice the enormous chair in the middle of that photograph. Whiskey Tango Foxtrot?
Yeah, I noticed it too, although my response was “Why are there mountains behind the Anacostia Big Chair?”
October 14th, 2009 at 7:52 am
I can’t believe I’m the first one to notice the enormous chair in the middle of that photograph.
Actually, Matt has used that picture at least once before in my recollection, and the chair was discussed on that occasion. So you are in fact late, not early, to the party.
October 14th, 2009 at 7:58 am
I suppose it’s possible to work out a couple of regimes where congestion pricing increases traffic. Either (1) demand for driving is highly price-sensitive but nobody is willing or able to accept an alternative or means to reduce the per passenger cost (few if any people adopt mass transit, carpooling, slug lines, telecommuting, et cetera) so to save a few bucks people divert onto non-priced roads and get stuck in horrible hour-long traffic jams while the congestion-priced roads are empty, or (2) we switch to a true 24-hour society, so that instead of having three traditional shifts with the 9-5 (or 8-5 or 9-6 for people who don’t get paid lunch hours) shift accounting for the large majority of workers, we get 24 shifts of equal size (for workers and for students alike), so that demand for roadway space is equal at 9:00 am and 9:00 pm and 3:00 am and so on. If use is spread out instead of clustering around certain key times, there might be more overall use even though there’s far fewer traffic jams.
Version (2) does not strike me as plausible. Version (1) would seem to be contraindicated by rational markets theory (very few people value their time at less than a couple of bucks per hour, and those that do generally can’t afford cars) but given the limits of rationality I suppose it can’t be entirely ruled out. However, if it develops it’s easily dealt with — change the pricing level or the pricing area or the flow pattern of the other streets.
October 14th, 2009 at 8:35 am
I’m with 11.
He is actually making the Yogi Berra argument. Amazing.
October 14th, 2009 at 8:58 am
[...] More Reasons Why David Owen Is Wrong on Congestion Pricing (Streetsblog Cap Hill, Yglesias) [...]
October 14th, 2009 at 9:52 am
For Jason L (and others) – one part of Owen’s article that nobody is mentioning is the reference to Ramp Metering. This is one situation where a reduction in traffic actually led to an INCREASE in the number of cars. So sitations like this do exist…
October 14th, 2009 at 10:08 am
@27: Poor people already can’t use roads, no matter how many more of them you build. If a soup spoon cost $20,000 new and at least $2,000 used, plus maintenance costs, fuel and mandatory insurance, then liberals might have a different attitude towards just opening another soup kitchen, too. (And Uri Geller would make a killing off the Cash for Bent Spoons program.)
Roads are sort of a public good, but because of the entry barrier to using them, they’re a peculiar one.
October 14th, 2009 at 10:40 am
The reason ramp meters can increase traffic volumes on the relevant highway in some circumstances is that they can actually work through two different mechanisms: diverting traffic through additional queue rationing at the point of entry, or simply breaking up “platoons” of cars that can lead to additional congestion.
The first mechanism logically should not lead to increased traffic volumes. But the second mechanism might insofar as it actually reduces the amount of total congestion, and thus total queue rationing, associated with a given volume of traffic.
As we have discussed above, congestion pricing per is just a way of substituting another form of rationing for queue rationing, so shouldn’t have a volume-increasing effect. On the other hand, some sort of widespread traffic coordination mechanism might have that effect through reducing congestion for a given total volume, but such a mechanism isn’t inherent to congestion pricing.
October 15th, 2009 at 12:55 am
An anecdote for the start: recently I have seen a reverse of “nobody goes to this restaurant anymore…”: in a hiking blog someone claimed that a certain new trail is “increasingly popular”, and a follow-up poster wrote “yes, this place is absolutely great! I hiked there for two days and did not meat anyone!”. I guess that “Allegheny Front Trail” is safe from congestion pricing for a while
In any case, it was discussed that proper pricing of street parking and congestion fees reduces idling and futile driving. Moreover, I am not sure if this is “regressive”. Low income employees carpool much more readily than high income — if they have no good access to transportation. Lone obstinate drivers would be disproportionally higher middle class.